Final _year_Prediction_system_review.ppt

Final year project on PowerBi,AL,ML.

1. ABSTRACT In an era of global education, the demand for accurate and efficient methods to predict foreign university admissions is paramount. This project explores the application of machine learning algorithms to predict the likelihood of admission based on key features such as GRE Score, TOEFL score, University rating, Statement of Purpose (SOP), Letter of Recommendation (LOR), CGPA, and Research experience. The dataset, comprising a diverse set of applicants' profiles, serves as the foundation for model development and evaluation. Various machine learning algorithms, including XGBoost , Decision Tree, Random Forest, Gradient Boosting, k-nearest Neighbors (KNN), Linear Regression, and Support Vector Machine (SVM), were employed to identify the most effective model for predicting admission outcomes. Ultimately, the project not only offers a practical solution for foreign university admission prediction but also underscores the transformative impact of machine learning in optimizing critical decision-making processes in education.

2. INTRODUCTION: In the dynamic landscape of global education, the process of foreign university admissions stands as a critical gateway for aspiring students and a complex decision-making challenge for institutions. The aim of this project is to harness the power of machine learning to streamline and optimize this intricate process. By leveraging a dataset enriched with key applicant attributes such as GRE Score, TOEFL score, University rating, SOP, LOR, CGPA, and Research experience, the project seeks to develop a predictive model for foreign university admission outcomes. This project holds significant implications for both prospective students and academic institutions. For applicants, it provides a tool for assessing their chances of admission based on quantifiable metrics, fostering informed decision-making in the pursuit of higher education abroad. Meanwhile, universities stand to benefit from a more streamlined and data-driven admissions process, optimizing resources and enhancing the overall efficiency of their academic intake procedures. The journey of this project involves not only the development of a high-accuracy predictive model but also an insightful comparison of various machine learning techniques

3. EXISTING SYSTEM & DISADVANTAGES: The current system for foreign university admissions relies predominantly on manual processes and human judgment, involving extensive reviews of application materials such as GRE scores, TOEFL scores, and recommendation letters. Admission decisions are often subjective, varying across institutions and lacking a standardized framework. Prospective students submit their credentials, and selection committees manually assess these materials to make admission decisions. This manual approach is time-consuming, prone to errors, and can introduce biases. DISADVANTAGES: Resource Intensiveness Inconsistency Across Institutions Inefficiency in Handling Large Volumes Neglect of Diverse Data Limited Predictive Power Subjectivity and Bias

4. PROPOSED SYSTEM & ADVANTAGES: The proposed system aims to overcome the limitations of the existing manual admission process by leveraging machine learning for foreign university admissions. The proposed system make use of automatic application evaluation,algorithm selection & optimization, transparent decision making to ensure security for the applicants data ADVANTAGES: Increased Accuracy Objective Decision-Making Efficiency and Time Savings Versatility with Multiple Algorithms Empowered Decision-Making for Applicants Transparency in Evaluation

5. Requirements Analysis: HARDWARE REQUIREMENTS

6. SOFTWARE REQUIREMENTS

7. Modules: Data Ingestion : Gets data from different sources like files or databases . 2 . Data Preprocessing : Cleans and organizes the data for analysis by handling missing value and converting formats . 3 . Feature Engineering : Creates new data features or modifies existing ones to improve the model's accuracy. 4. Model Training : Teaches the model patterns in the data to make predictions . 5. Model Evaluation : Checks how well the model performs using specific measures .

8. 6. Interpretability : Helps understand why the model makes certain predictions. 7. User Interface : Builds the part of the system that users interact with. 8. Deployment : Gets the model ready to use in real-world applications . 9 . Logging and Monitoring : Keeps track of how the system performs and any user interactions . 10. Security : Protects user data and the system from potential threats.

9. Technologies Used:

10. Implemented Models

11. System design:

12. Output screens: Home Page Login Page

13. Register page User dashboard

14. My profile Predict page

15. Graph Comparision predict result page

16. Test cases

17. Conclusion: This project uses machine learning to predict foreign university admissions, making the process fairer and more efficient. By analyzing applicant data like test scores, recommendations, and GPA, it built a predictive model using Linear Regression. With 81% accuracy, it helps students understand their admission chances and universities improve their selection process. The user-friendly interface benefits students and institutions alike, reducing bias and subjectivity. Future improvements could include advanced algorithms and more features, advancing data-driven decision-making in global education. This project showcases the power of machine learning in transforming university admissions.

18. References: 1. Sridhar et al. (2020) developed a University Admission Prediction System using Stacked Ensemble Learning. 2. Haythorhwaithe et al. (2013) introduced a special issue on learning analytics. 3. Ragab et al. (2012) proposed HRSPCA, a Hybrid Recommender System for predicting college admission. 4. Sivasangari et al. (2021) worked on predicting the probability of university admission using Machine Learning. 5. Türker et al. (2020) presented a Deep Hybrid Recommender System. 6. Pandian ( 2019) reviewed machine learning techniques for managing voluminous information. 7. Kumar (2021) constructed a Hybrid Deep Learning Model for predicting children's behavior based on their emotional reactions.